SUMMARY Alcohol use disorder (AUD) affects more than 12% of the US population and is the fourth leading preventable cause of death in the US. AUD is associated with compulsive drinking and emergence of a negative emotional state during abstinence, resulting in two-thirds of afflicted people relapsing within months of completing alcohol cessation treatment. Repeated cycles of intoxication and abstinence are linked to persistent alterations in the brain that remain long after detoxification, which drive relapse of drinking, often at levels higher than before abstinence. Identifying the molecular mechanisms that lead to hazardous alcohol use and contribute to the high rates of relapse is imperative for the effective design of better treatments to curb alcohol abuse. Our previous studies illustrate the power of integrating epigenetic, transcriptomic, circuitry, pharmacology and behavioral tools to elucidate the molecular underpinnings contributing to hazardous alcohol use. Those studies, using genome- wide DNA methylation (DNAm) analysis of the nucleus accumbens (NAc) to compare macaque alcohol-naive and alcohol-drinkers, identified differential DNAm (D-DNAm) signals, mapping to genes not previously linked to alcohol use, but with high relevance to synaptic plasticity modulation. Experimental manipulation of two of these DNAm-linked genes resulted in altered glutamate and GABA neurotransmission and changes in ethanol intake. The the next critical step is to understand how repeated abstinence/relapse cycles alters DNAm signals and how these changes contribute to adaptations in neurocircuitry. We hypothesize that some D-DNAm signals associated with chronic alcohol use may persist during abstinence, contributing to risk of relapse. In addition, de novo D-DNAm signals generated during repeated cycles of abstinence/relapse, may further heighten relapse risk. To identify abstinence-associated DNAm, we will use the highly translational macaque alcohol self- administration model and our proven genome-wide approach to identify D-DNAm in the NAc of macaques following > 12 months of chronic alcohol self-administration and then 3 cycles of forced abstinence (each cycle: 1 month abstinence and 3 months of open-access). After integrating gene and transcript variant expression and D-DNAm data generated from the same subjects and tissues, a subset of compelling, novel targets will be selected for functional study using pharmacological and genetic manipulation approaches in a mouse chronic intermittent ethanol (CIE) model. This rodent CIE model has been extensively used to model cycles of alcohol exposure and withdrawal leading to an escalation of drinking during relapse. In addition to recording ethanol intake, the effects of target manipulation on neurotransmission will be evaluated using patch-clamp electrophysiological analysis. Overall, these studies will aid in our understanding of the molecular mechanism(s) establishing risk for relapse. This information will be critical to advancing the development of effective therapies for alcohol abuse and to prevent relapse.